Python SDK

⚙️ PageIndex Python SDK

Submit documents for PageIndex Processing

Chat with your PageIndex-processed documents

Initialize

To get started, visit the PageIndex Developer Dashboard and generate your API key .


pip install -U pageindex


from pageindex import PageIndexClient
 
pi_client = PageIndexClient(api_key="YOUR_API_KEY")

Submit a Document

Upload a PDF for processing. The returned doc_id is used for subsequent operations.


result = pi_client.submit_document("./2023-annual-report.pdf")
doc_id = result["doc_id"]

Check if processing has completed:


status = pi_client.get_document(doc_id)["status"]
if status == "completed":
    print('Document processing completed')

See Document Processing for full parameters and response format.

Chat API (beta)

Ask questions about one or more processed documents. The Chat API runs a full agentic, reasoning-based RAG workflow under the hood using PageIndex.

Chat with a specific document


response = pi_client.chat_completions(
    messages=[{"role": "user", "content": "What are the key findings in this document?"}],
    doc_id="pi-abc123def456"
)
 
print(response["choices"][0]["message"]["content"])

Chat with multiple documents


response = pi_client.chat_completions(
    messages=[{"role": "user", "content": "Compare these two documents on the results"}],
    doc_id=["pi-abc123def456", "pi-abc123ghi789"]
)
 
print(response["choices"][0]["message"]["content"])

See Chat API (beta) for the full reference.

💡

Want results instantly with streaming? The Chat API also supports streaming responses.

You can build a customized agentic retrieval API by prompting the Chat API. See this notebook for an example.

Get Document Tree Index

Get the hierarchical tree index generated for a processed document.


tree_result = pi_client.get_tree(doc_id)["result"]

See Document Processing for more details on the tree structure.

SDK Full Reference

Document Processing: Submit documents for PageIndex processing.
Chat API (beta): Chat with your PageIndex-processed documents.

Also see OCR for structure-preserving PDF to markdown conversion.